Bug report #10396

Raster rendering in 2.3 vs 2.2

Added by andskog - over 10 years ago. Updated over 5 years ago.

Status:Closed
Priority:Normal
Assignee:-
Category:Map Canvas
Affected QGIS version:master Regression?:No
Operating System: Easy fix?:No
Pull Request or Patch supplied:No Resolution:end of life
Crashes QGIS or corrupts data:No Copied to github as #:18813

Description

The raster zooming and panning behavior seems to be different in master than the current 2.2 version. Although not bugs, maybe not improvements from 2.2 to 2.3.

When zooming in, 2.2 seems to stay at the current level while preparing the new zoomed in map, and then zooms in when ready.
Master (2.3) seems to zoom in, blank the map, and render the new zoomed in map when ready. The 2.3 behavior is less eye pleasing.

Additionally, 2.2 renders the data much faster at my desk.

qg22rastertest.qgs (4.54 KB) andskog -, 2014-06-01 12:46 PM

qg23rastertest.qgs (4.64 KB) andskog -, 2014-06-01 12:46 PM

raster.mp4 (1.83 MB) Giovanni Manghi, 2014-06-03 04:48 AM

22.mp4 (252 KB) andskog -, 2014-06-03 02:11 PM

23.mp4 (197 KB) andskog -, 2014-06-03 02:11 PM

History

#1 Updated by Martin Dobias over 10 years ago

The "old zoomed map -> blank -> new map" is not something specific to raster layers, it may happen also with vectors, it is just more obvious with rasters. The behaviour will need some updates to make better impression. The reason for it is that first 250 ms, the canvas shows previously rendered map transformed to the new extent (e.g. zoomed in). Then, the old map is replaced with preview of newly rendered map - which may be still blank at that point, giving the resulting effect. In the meantime you could increase the refresh ration in Options to a greater value to remove the effect.

Regarding faster/slower rendering, could you provide approximate timings of 2.2 vs 2.3 ? Also, we would need the data you used for the comparison.

#2 Updated by Giovanni Manghi over 10 years ago

  • Status changed from Open to Feedback

#3 Updated by andskog - over 10 years ago

Rendering speed:

Used the DEM+HDF files available at http://dds.cr.usgs.gov/srtm/version2_1/SRTM30/w020n90/
Set project to EPGS 32661 (enabled on-the-fly projection), setting one extent covering the whole dataset in the left part of the canvas, another extent to the right part. Then using Zoom last / Zoom next for repeated tests.

Produced these averages:
QGIS2.2: 2.7 sec.
QGSI2.3: 4.3-5.3, most results above 5 sec.

#4 Updated by Giovanni Manghi over 10 years ago

Then using Zoom last / Zoom next for repeated tests.

here (Linux) it takes no time in zooming_in_out/last/next on both 2.2 and master

#5 Updated by Etienne Tourigny over 10 years ago

maybe it's due to different settings, make sure they bot have rendering cache turned on (or off)

#6 Updated by andskog - over 10 years ago

It's evident with any raster here. And regardless of map update interval settings. Cache settings are the same in both 2.2 and 2.3 (off). (It also reproduces when on-the-fly reprojection is turned off) ...

#7 Updated by Giovanni Manghi over 10 years ago

andskog - wrote:

It's evident with any raster here. And regardless of map update interval settings. Cache settings are the same in both 2.2 and 2.3 (off). (It also reproduces when on-the-fly reprojection is turned off) ...

then better to attach here a full project and/or a screencast (to try better understand your issue).

#8 Updated by andskog - over 10 years ago

I've attached QGIS 2.2 and 2.3 project files. Makes less sense to attach small (<5MB) raster file, as you (seems like it) have more powerful computers. But the settings should all be there. (Tested with ETOPO1 raster, available here: http://www.ngdc.noaa.gov/mgg/global/relief/ETOPO1/data/ice_surface/grid_registered/georeferenced_tiff/ETOPO1_Ice_g_geotiff.zip)

#9 Updated by Giovanni Manghi over 10 years ago

andskog - wrote:

I've attached QGIS 2.2 and 2.3 project files. Makes less sense to attach small (<5MB) raster file, as you (seems like it) have more powerful computers. But the settings should all be there. (Tested with ETOPO1 raster, available here: http://www.ngdc.noaa.gov/mgg/global/relief/ETOPO1/data/ice_surface/grid_registered/georeferenced_tiff/ETOPO1_Ice_g_geotiff.zip)

see attached screencast

#10 Updated by andskog - over 10 years ago

To me it seems kind of apparent in your screen capture too. I've attached mine. 2.2 takes ~1.3s, 2.3 takes ~2.3s. (The difference is there also when zooming in on places not zoomed in to before.)

#11 Updated by Giovanni Manghi over 10 years ago

  • Status changed from Feedback to Open

#12 Updated by Jürgen Fischer over 10 years ago

  • Target version changed from Version 2.4 to Future Release - High Priority

#13 Updated by Giovanni Manghi over 7 years ago

  • Regression? set to No
  • Easy fix? set to No

#14 Updated by Giovanni Manghi over 5 years ago

  • Resolution set to end of life
  • Status changed from Open to Closed

Also available in: Atom PDF